Particulate Emissions from Apartment House Boilers and Incinerators

The body of information presented in this paper is directed to researchers in stack testing methodology and to those concerned with reduction of emissions through equipment upgrading programs. Extensive testing was done using the U.S. Environmental Protection Agency’s Method 5 stack sampling train to obtain emission factors for existing apartment house boilers and incinerators in the City of New York. In addition to calculating emission factors, stack emission data were examined to compare results of simultaneous emission tests and to compare the dry particulate catch of the sampling train with the total particulate catch which included the impinger catch. Conclusions reached as a result of the testing were that published emission factors for boilers burning moderately high-sulfur residual oil are applicable to New York City boilers burning low-sulfur residual oil. In addition, it was found that the back half of the sampling train—the impinger section—collects a relatively constant amount of material when sampling oil-fired boilers. This may be due to absorption of S0₂ and S0₃ in the impingers and the subsequent formation of sulfuric acid. Comparison of simultaneous boiler tests indicated that the sampling train may be sensitive to variations in operating personnel, sampling conditions, and boiler operation. From tests of on-site incinerators, it was determined that previously published emission factors may be too high for well maintained and properly operated incinerators. The back half particulate catch was found to be relatively large which may have resulted from condensation of unburned organics from the burning waste material.     

Hot Filter/Impinger and Dilution Sampling for Fine Particulate Matter Characterization from Ferrous Metal Casting Processes

Abstract

A study using two stack-sampling methodologies for collecting particulate matter (PM) emissions was conducted using a hot filter followed by a cold impinger sampling train and a dilution sampler. Samples were collected from ferrous iron metal casting processes that included pouring molten iron into a sand mold containing an organic binder, metal cooling, removal of the sand from the cooled casting (shakeout), and postshakeout cooling. The shakeout process contributed more to PM emissions than the metal pouring and cooling processes. Particulate matter less than 2.5 μm in aerodynamic diameter (PM_2.5) mass emissions for the entire casting cycle ranged from 3.4 to 4.7 lb/t of metal for the hot filter/impinger method and from 0.8 to 1.8 lb/t of metal for the dilution method. Most of the difference was due to PM captured by the impingers, much of which was probably dissolved gases rather than condensable vapors. Of the PM fraction captured by the impingers, 96–98% was organic in nature. The impinger PM fraction contributed 32–38% to the total suspended particle mass and caused a factor of 2–4 positive bias for PM_2.5 emissions. For the pouring and cooling processes only, the factor increased to over seven times.     

Hot filter/impinger and dilution sampling for fine particulate matter characterization from ferrous metal casting processes

A study using two stack-sampling methodologies for collecting particulate matter (PM) emissions was conducted using a hot filter followed by a cold impinger sampling train and a dilution sampler. Samples were collected from ferrous iron metal casting processes that included pouring molten iron into a sand mold containing an organic binder, metal cooling, removal of the sand from the cooled casting (shakeout), and postshakeout cooling. The shakeout process contributed more to PM emissions than the metal pouring and cooling processes. Particulate matter less than 2.5 microm in aerodynamic diameter (PM2.5) mass emissions for the entire casting cycle ranged from 3.4 to 4.7 lb/t of metal for the hot filter/impinger method and from 0.8 to 1.8 lb/t of metal for the dilution method. Most of the difference was due to PM captured by the impingers, much of which was probably dissolved gases rather than condensable vapors. Of the PM fraction captured by the impingers, 96-98% was organic in nature. The impinger PM fraction contributed 32-38% to the total suspended particle mass and caused a factor of 2-4 positive bias for PM2.5 emissions. For the pouring and cooling processes only, the factor increased to over seven times.     

Sampling and analysis of biological aerosols

The extreme particle size range and enormous heterogeneity of airborne biological particles make sampling a significant challenge. Three major sampler types available include gravity devices, impactors and suction samplers. Gravity methods, while most commonly used, are neither qualitatively or quantitatively accurate and of very limited use. Impaction samplers (rotating, centrifugal) accelerate air by rotating the collecting surface or with a fan. Particles are collected from measured volumes of air but these devices preferentially sample particles larger than 10 μm. Suction samplers, which efficiently collect particles of a wide size range from measured volumes of air, include slit samplers, cascade impactors, filtration devices and liquid impingers. Suction samplers can retrieve viable particles by direct impaction on culture media, or by subsequent culture of impinger fluid or filter eluates. Nonviable particles can often be identified by microscopic examination of slides, filters or filtrates of impinger fluids. Immunoassays and biochemical assays can be used with impinger fluid and filter eluates to assess antigen and toxin levels in measured air samples.     

Issues Related to Solution Chemistry in Mercury Sampling Impingers

ABSTRACT

Analysis of Hg speciation in combustion flue gases is often accomplished in standardized sampling trains in which the sample is passed sequentially through a series of aqueous solutions to capture and separate oxidized Hg (Hg²⁺) and elemental Hg (Hg⁰). Such methods include the Ontario Hydro (OH) and the Alkaline Mercury Speciation (AMS) methods, which were investigated in the laboratory to determine whether the presence of Cl₂ and other common flue gas species can bias the partitioning of Hg⁰ to front impingers intended to isolate Hg²⁺ species. Using only a single impinger to represent the front three impingers for each method, it was found that as little as 1-ppm Cl₂ in a simulated flue gas mixture led to a bias of approximately 10-20% of Hg⁰ misreported as Hg²+ for both the OH and the AMS methods. Experiments using 100-ppm Cl₂ led to a similar bias in the OH method, but to a 30-60% bias in the AMS method. These false readings are shown to be due to liquid-phase chemistry in the impinger solutions, and not necessarily to the gas-phase reactions between Cl₂ and Hg as previously proposed. The pertinent solution chemistry causing the interference     

Particle-Molecule Collection by Sonic Flow Impingers

The information presented in this paper is directed to those industrialists and researchers interested in molecule and near micron and submicron particle sampling by a convenient, inexpensive, and sufficiently accurate method. The use of two sonic flow impingers in series is predicted to collect 98-99% of a phosphoric acid aerosol having a mass median diameter of 0.7 p. The first impinger of a new design is shown, in field sampling, to collect simultaneously 90-98% of the aerosol and 95% of molecular fluoride compounds. On the same aerosol, the standard Greenburg-Smith impinger shows a low collection performance, less than 50% at 1 cfm and 65% at sonic velocity. The use of a first impinger at sonic flow greatly simplifies sampling procedures by eliminating the need for a test meter and associated pressure and temperature measurements. Also, for small particle sizes isokinetic sampling is not necessary and sampling line losses by deposition are shown to be less than 1%. The low initial cost of the sampling units and the low manpower requirements for setting up and supervising sampling make possible the taking of a number of samples sufficient to establish emissions over extended intervals of time. The equipment is especially suited to locations having multiple emission sources, or for locations requiring simultaneous sampling of many points.     

Woodstove Smoke and CO Emissions: Comparison of Reference Methods with the VPI Sampler

A new field sampler has been developed for measuring the particulate matter (PM) and carbon monoxide emissions of woodburning stoves. Particulate matter is determined by carbon balance and the workup of a sample train which is similar to a room-temperature EPA Method 5G train. A steel tank, initially evacuated, serves as the motive force for sampling and also accumulates a gas sample for post-test analysis of time-averaged stack CO and CO₂ concentrations. Workup procedures can be completed, within 72 hours of sampler retrieval. The system has been compared to reference methods in two laboratory test series involving six different woodburning appliances and two independent laboratories. The correlation of field sampler emission rates and reference method rates is strong.     

Trace Metal Samples Collected in the Front and Back Halves of the E.P.A. Stack Sampling Train

A Purdue University industrial source sampling team has been involved since 1972 with a number of industrial and municipal collaborators, in order to characterize the flow of trace metals into the atmosphere. The plants involved in this cooperative research effort include the East Chicago municipal incinerator, rated at 450 ton/day of residential and commercial solid waste; a multiple furnace open hearth shop at a Northwest Indiana steel mill producing approximately 8 million ton/yr of steel; a coker arid sinter plant serving a 100,000 ton/yr vertical retort zinc production facility; and the Purdue University coal fired power plant equipped with 250,000 lb/hr steam boilers. At each of these facilities a number of stack samples have been obtained using the standard E.P.A. train. Analysis of the probe, filter, and impinger catch showed that in each case the front half of the E.P.A. Method 5¹ sampling train was highly efficient for collection of trace metal particulate.     

Ozone Measurements in South Carolina Using Passive Samplers

Abstract

Passive samplers with two different collection substrates were used to obtain an average ozone concentration for 1 month during the summer of 2002 for each South Carolina county. One sampler contained a filter coated with indigo carmine, whose color fades when exposed to ozone. The fading was measured by reflectance spectroscopy. The other sampler contained filters that were coated with nitrite, which is oxidized to nitrate when exposed to ozone. The nitrate was measured by ion chromatography.

Calibration curves were developed for the two methods by comparing color fading from indigo carmine and nitrate ion concentration from the nitrite filter with ambient ozone concentration measured by a co-located reference continuous UV ozone analyzer. These curves were used to calculate integrated ozone concentrations for samplers distributed across South Carolina.

Using the indigo carmine method, the average ozone concentrations ranged from 21 to 64 ppb (average = 46 ± 7.9 ppb, n = 58) across the 46 counties in the state during one summer month of 2002. Concentrations for the same time period from the nitrite-coated filters ranged from 23 to 62 ppb (average = 41 ± 8.1 ppb, n = 58). Also for the same time period, the 23 continuous UV photometric ozone monitors operated by the South Carolina Department of Health and Environmental Control at sites within 10 miles of some of the passive monitors showed ozone concentrations ranging from 28 to 50 ppb (average = 39 ± 6.3 ppb, n = 22).     

Rationale for an Eight-Hour Ozone Standard

As part of the exposure assessment scheme for a community-based air pollution health effects study, 43 homes of study participants, located in two Houston neighborhoods, were monitored for weekly-average indoor formaldehyde levels by means of diffusion samplers. Consecutive 12-hour aldehyde sampling for one-week periods was conducted in 12 of the homes by means of pumps and impingers. In six houses where simultaneous monitoring with both methods occurred, good correlation between the results from the diffusion samplers and the standard impinger method was observed. Diffusion sampler precision was variable and lower than expected, and a small positive measurement bias could be inferred. The distribution of house-average indoor formaldehyde concentrations from diffusion monitoring was similar to that obtained during a previous housing survey in Houston, with concentrations in 19 percent of the homes exceeding 0.10 ppm. Formaldehyde levels in this group of conventional, mostly older homes could not be associated with smoking, cooking, home age or structure type. However, there was a statistically significant difference between mean indoor concentrations in the two neighborhoods.     

Field performance evaluation during fog-dominated wintertime of a newly developed denuder-equipped PM1 sampler

This study presents the performance evaluation of a novel denuder-equipped PM₁ (particles having aerodynamic diameter less than 1 μm) sampler, tested during fog-dominated wintertime, in the city of Kanpur, India. One PM₁ sampler and one denuder-equipped PM₁ sampler were co-located to collect ambient PM₁ for 25 days. The mean PM₁ mass concentration measured on foggy days with the PM₁ sampler and the denuder-equipped PM₁ sampler was found to be 165.95 and 135.48 μg/m³, respectively. The mean PM₁ mass concentration measured on clear days with the PM₁ sampler and the denuder-equipped PM₁ sampler was observed to be 159.66 and 125.14 μg/m³, respectively. The mass concentration with denuder-fitted PM₁ sampler for both foggy and clear days was always found less than the PM₁ sampler. The same drift was observed in the concentrations of water-soluble ions and water-soluble organic carbon (WSOC). Moreover, it was observed that the use of denuder leads to a significant reduction in the PM positive artifact. The difference in the concentration of chemical species obtained by two samplers indicates that the PM₁ sampler without denuder had overestimated the concentrations of chemical species in a worst-case scenario by almost 40 %. Denuder-fitted PM₁ sampler can serve as a useful sampling tool in estimating the true values for nitrate, ammonium, potassium, sodium and WSOC present in the ambient PM.     

Comparison of variability in dioxin and furan data acquired using single train and simultaneous multiple train stack sampling methods

The variance of polychlorinated dibenzo-p-dioxin (PCDD; dioxin) and polychlorinated dibenzofuran (PCDF; furan) data obtained from single- and simultaneous multiple train methods was compared. Single train triplicate data were used from 4 stack tests obtained from a long dry kiln cement plant and 18 stack tests from a municipal solid waste (MSW) incinerator. Data from the American Society of Mechanical Engineers (ASME) report Reference Method Accuracy and Precision (ReMAP) (Lanier and Hendrix, 2001) were used for the simultaneous multiple samples, which accounted for 27 data points. Nineteen data points were acquired from an ASME research facility, 5 from a MSW incinerator unrelated to the single train MSW incinerator, and 3 from a lightweight aggregate kiln (LWAK). The ReMAP procedure was used to determine the relationship between the standard deviation and the concentration of the single train and simultaneous multiple train data. Results indicated that there was benefit from the use of simultaneous multiple train sampling for concentrations above 129 pg toxic equivalency (TEQ)/m³. There was no indication of benefit from the use of simultaneous multiple train sampling at concentrations below 129 pg TEQ/m³.

Implications:Precision of stack sampling data can be the difference between meeting and failing compliance limits. Generally, three dioxin/furan samples are acquired when stack sampling to meet compliance regulations. A reliable estimation of the data’s true concentration is not possible with this small amount of data. Increasing the precision decreases the chance that the acquired concentration deviates greatly from the true concentration. Facilities that use the appropriate stack sampling method will benefit by either improved data precision or minimal stack sampling expenses. The observations made suggest that facilities that are expected to have dioxin/furan concentrations above 129 pg TEQ/m³ would increase the precision of samples by using simultaneous multiple train sampling.     

Measurement of particulate aliphatic and polynuclear aromatic hydrocarbons in Santiago de Chile: source reconciliation and evaluation of sampling artifacts

Using a novel sampler, particulate organic compounds were collected in Santiago de Chile from June 9 to August 10, 1997. This sampler consists of a diffusion denuder to remove gas-phase organics prior to particle collection, a Teflon filter, and a PUF cartridge downstream of the filter. PAHs and n-alkanes were measured using gas chromatography/mass spectrometry analysis. Volatilization of particles collected on the Teflon filter varied from 15 to 85% for both n-alkanes and PAHs, with strong dependence on molecular weight. The relative distribution of n-alkanes and the values of molecular diagnostic ratios, such as Carbon preference index, indicated a mixed origin with strong anthropogenic input. Indeed, CPI values ranged from 0.66 to 1.96 (for the whole range of n-alkanes). The percent contribution of leaf “wax” n-alkanes (4.55–20.83%) indicated the low contribution of biogenic sources. In addition, the distribution pattern of PAHs was characteristic of anthropogenic emissions. The dominant contribution of combustion-related PAHs (CPAHs), 74–84%, indicated that vehicular emissions was the major source of PAHs.     

Issues related to solution chemistry in mercury sampling impingers

Analysis of Hg speciation in combustion flue gases is often accomplished in standardized sampling trains in which the sample is passed sequentially through a series of aqueous solutions to capture and separate oxidized Hg (Hg2+) and elemental Hg (Hg0). Such methods include the Ontario Hydro (OH) and the Alkaline Mercury Speciation (AMS) methods, which were investigated in the laboratory to determine whether the presence of Cl2 and other common flue gas species can bias the partitioning of Hg0 to front impingers intended to isolate Hg2+ species. Using only a single impinger to represent the front three impingers for each method, it was found that as little as 1-ppm Cl2 in a simulated flue gas mixture led to a bias of approximately 10-20% of Hg0 misreported as Hg2+ for both the OH and the AMS methods. Experiments using 100-ppm Cl2 led to a similar bias in the OH method, but to a 30-60% bias in the AMS method. These false readings are shown to be due to liquid-phase chemistry in the impinger solutions, and not necessarily to the gas-phase reactions between Cl2 and Hg as previously proposed. The pertinent solution chemistry causing the interference involves the hypochlorite ion (OCl-), which oxidizes Hg0 to soluble Hg2+. Addition of sodium thiosulfate (Na2S2O3) to the front impinger solutions eliminates this false positive measurement of Hg2+ by selectively reacting with the OCl- ion. In general, the presence of SO2 also mitigates this interference in the same way, and so this bias is not likely to be a factor for Hg speciation measurements from actual coal combustion flue gases. It might, however, be a problem for those few combustor flue gas measurements and research studies where Cl2 is present without appreciable amounts of SO2.     

Integrated Sampling and Analysis Methods for Determining NO x Emissions at Electric Utility Plants

Two integrated sampling and analysis methods for determining NO_x emissions in electric utility plants were developed and field tested. The collection systems consist of: a 4.0% potassium permanganate-2.0% sodium hydroxide solution in restricted-orifice impingers, and a 5A° molecular sieve in midget impingers. Sample analysis is accomplished by a colorimetric or ion-chromatographic procedure with the alkaline-permanganate method and by a colorimetric procedure with the molecular sieve method. The alkalinepermanganate method gives excellent agreement with the EPA reference method, Method 7, for NO _x measurements. The molecular sieve method shows a significant negative bias relative to Method 7. It is anticipated that the permanganate methods will be proposed as alternates to Method 7, for NO _x determinations, under the EPA New Source Performance Standards.     

The Determination of Formaldehyde,Acrolein, and Low Molecular Weight Aldehydes in Industrial Emissions on a Single Collection Sample

Formaldehyde, acrolein, and low molecular weight aldehydes are collected in 1% NaHSO₃ solution in Greenberg-Smith or midget impingers from industrial effluents. Efficiency of collection is excellent when two impingers are used in series. Formaldehyde is measured in an aliquot of the collection medium by the chromo-tropic acid procedure, acrolein by a modified mercuric-chloride-hexylresorcinol procedure, and C₂-C₅ aldehydes by a gas chromatographic procedure. The method permits the analysis of all C₁-C₅ aldehydes on a single collection sample. Data on aldehyde concentrations from a variety of effluents are presented.     

A Transition-Flow Reactor Tube for Measuring Trace Gas Concentrations

Dry deposition contributes significantly to the acidification of ecosystems. However, difficulties in measuring dry deposition of reactive gases and fine particles make routine direct monitoring impractical. An alternate approach is to use the “concentration monitoring” method in which dry deposition flux is estimated as the product of measured concentration and estimated deposition velocity. A sampling system that performs over the period of 6 hours to 7 days, depending on atmospheric concentrations, has been developed. It consists of a Teflon cyclone to exclude particles larger than about 2 μm, selective solid adsorption media for reactive gases—some of which are sampled from a transition flow to avoid possible bias from particle evaporation, a particle filter, and a final gas adsorption filter to collect the remaining trace gas. The sampler Is the first reported application of transition flow mass transfer for the collection and quantitative measurement of trace atmospheric gases. Laboratory and field tests have shown that the sampler performs well for HNO₃(g).     

Woodstove smoke and CO emissions: comparison of reference methods with the VIP sampler

A new field sampler has been developed for measuring the particulate matter (PM) and carbon monoxide emissions of woodburning stoves. Particulate matter is determined by carbon balance and the workup of a sample train which is similar to a room-temperature EPA Method 5G train. A steel tank, initially evacuated, serves as the motive force for sampling and also accumulates a gas sample for post-test analysis of time-averaged stack CO and CO2 concentrations. Workup procedures can be completed within 72 hours of sampler retrieval. The system has been compared to reference methods in two laboratory test series involving six different woodburning appliances and two independent laboratories. The correlation of field sampler emission rates and reference method rates is strong.     

South Philadelphia passive sampler and sensor study

From June 2013 to March 2015, in total 41 passive sampler deployments of 2 wk duration each were conducted at 17 sites in South Philadelphia, PA, with results for benzene discussed here. Complementary time-resolved measurements with lower cost prototype fenceline sensors and an open-path ultraviolet differential optical absorption spectrometer were also conducted. Minimum passive sampler benzene concentrations for each sampling period ranged from 0.08 ppbv to 0.65 ppbv, with a mean of 0.25 ppbv, and were negatively correlated with ambient temperature (–0.01 ppbv/°C, R² = 0.68). Co-deployed duplicate passive sampler pairs (N = 609) demonstrated good precision with an average and maximum percent difference of 1.5% and 34%, respectively. A group of passive samplers located within 50 m of a refinery fenceline had a study mean benzene concentration of 1.22 ppbv, whereas a group of samplers located in communities >1 km distant from facilities had a mean of 0.29 ppbv. The difference in the means of these groups was statistically significant at the 95% confidence level (p < 0.001). A decreasing gradient in benzene concentrations moving away from the facilities was observed, as was a significant period-to-period variation. The highest recorded 2-wk average benzene concentration for the fenceline group was 3.11 ppbv. During this period, time-resolved data from the prototype sensors and the open-path spectrometer detected a benzene signal from the west on one day in particular, with the highest 5-min path-averaged benzene concentration measured at 24 ppbv.

Implications: Using a variation of EPA’s passive sampler refinery fenceline monitoring method, coupled with time-resolved measurements, a multiyear study in South Philadelphia informed benzene concentrations near facilities and in communities. The combination of measurement strategies can assist facilities in identification and mitigation of emissions from fugitive sources and improve information on air quality complex air sheds.